Virtual Dressing Room

Project Title

Virtual Dressing Room

Project Area of Specialization Artificial IntelligenceProject Summary

Garments industry is one of the top leading industries  worldwide .Current standards for trading in garments are wasting huge amount of time, not fulfilling customer’s requirements and hence not capable of catching more customers. Trying clothes in changing rooms is time consuming activity. Moreover this method cannot recognize and save the pattern of one’s choice as well as cannot provide any information about whatever is going in market regarding the dresses someone is trying on. So, the demand of a real time virtual reality dressing room is increasing with each day progressing towards a complete digital age. Some efforts regarding setting up such a Virtual Reality Dressing Room has been done in the past but they are either too inefficient to meet customer’s requirement (Like they display only 2D dresses) or they are highly expensive. Keeping these all things in mind, we have presented an efficient as well as cost effective Virtual Dressing Room in this paper that displays 3D dresses for its user and has been developed by using just a Microsoft’s Kinect360 Sensor and a set of software (Like unity 3D environment) to implement the logic. This system listens to almost all the basic requirements of a customer without costing them too high.

Project Objectives

Everyone has its own kind of interests in life. Making a virtual changing room was our dream as we are the first Pakistan students making it our final year project, because Virtual dressing mirrors are available with very high prices but we are introducing it in a very economical way. INNOVISION and SENSEMI have launched a Virtual dressing Mirror in Pakistan but having very high prices.

Project Implementation Method

Microsoft Kinect was specifically designed for detecting the humans in its field of view and capturing its movements which mainly requires locating certain data points on users. These data points known as joints are obtained through basic image processing within Kinect software. A total of 20 sub joints are located on a human body and this process is repeated 30 times a second. (i.e. 30fps) that enables motion capturing.

After defining dressing topology for 3D dresses, a user interface was to be designed for locating, instantiating and mapping these dresses to its user in camera’s field of view. For this, two camera object’s are defined inside the application that control the whole process:

• Main Camera
• Background Camera

Main camera object is used to display 3D dresses for user while background camera object is used to read information from Kinect sensor. Following components were added to these objects for making them work as per requirement

Benefits of the Project

1. You can try on clothes before buying them online. It shows you accurate fit and look.
2. As the buyers already know how the clothes will look and fit there is very low chance of return.
3. It keeps the customer engaged for a longer time, trying out different clothes on a look-a-like 3D avatar. This has more chances of conversion into a sale.
4. It saves you from the long queues at the malls or departmental stores.
5. It saves the retailer from damaged or soiled garments due to heavy try on by customers.
6. Customers can try their outfits anywhere, anytime and can compare looks of different apparels by taking the pictures
7. You don't have to go through wearing and taking off clothes again and again till you find the “one”.
8. It also provides information on available sizes and colors, and makes suggestions on matching garments to promote cross-sales.
9. Easy and quick conversion into 3D images of clothes included in the shop catalogue
10. The impact on consumer insight is also significant. In addition, because these high-tech approaches tend to increase overall satisfaction, they can become vehicles for turning one-time customers into repeat customers.

Technical Details of Final Deliverable

The technique used behind is  

Overlay Controller

It displays Kinect color camera feed in the background as well as positions and orients all other camera objects in the application to mach sensor’s point of view.

Category Selector

It filters the available Model Selector (categories) for detected user and allows model (dress) change in active category with hand swipes (left or right) and changes the category when a hand raise (left or right hand) is detected.

Model Selector

It locates the available model for the category, creates the category menu and instantiates the currently selected clothing model.

Final Deliverable of the Project Software SystemType of Industry IT Technologies Artificial Intelligence(AI)Sustainable Development Goals Industry, Innovation and InfrastructureRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	12000
Kinect Sensor	Equipment	1	4000	4000
LCD	Equipment	1	8000	8000